This project studies anisotropic visual sensitivity--the ability to see narrow contours on some orientations more clearly than on others, even when there is no optical error. Its specific aims are 1) to document the course of visual changes during childhood which eventually lead to greater sensitivity for horizontal and vertical contours than for diagonal ones in Western adults, and 2) to determine the physiological bases of and variables responsible for such changes. Three courses of investigation are proposed. One develops more efficient, criterion-free testing techniques for 4- to 6-year-old children. The second tests three hypotheses about what changes when adult eliminate anisotropy with practice, and investigates the specificity and persistence of sensory improvements resulting from practice. Third, variation in anisotropic sensitivity as a function of retinal locus is used to explore connections between visual performance and the architecture of the primary visual cortex. Two methods will be developed for testing young children. One is a two-alternative forced-choice (2AFC) paradigm incorporated into a computer-controlled video game, which would be useful in a wide range of developmental psychophysics research. The second method uses frequency-swept high-contrast sinusoidal gratings to evoke brain potentials (EPs). Pilot work shows that such a method is not only quick but is sensitive enough to pick up 5 c/deg differences in acuity as a function of orientation in adults. Signal detection, 2AFC and EP methods will be used to test three hypotheses about the source of sensory improvements resulting from practice--that practice 1) stabilizes accommodation 2) improves knowledge about where to look in orientation space for the stimulus, or 3) increases the number of visual cortical cells processing information on the practiced orientation. This research provides information about the plasticity of the adult human visual system, and may prove useful in devising therapies for meridional amblyopes. Finally, pilot work demonstrates that will stimuli scaled to the retinocortical magnification factor, variation in the pattern and magnitude of anisotropic sensitivity are associated with dimensions of eccentricity, left/right or nasal/temporal, and radial position. These findings will be extended and related to hypotheses about the inhomogeneity of cortical hypercolumns.